Update app.py

app.py

@@ -45,8 +45,7 @@ def pick_lang(langname):
     elif langname =="Faroese":
         df = datas.ds_f
 
-    dfd = df.data.to_pandas()
-    dfd = dfd.drop(columns=['audio', 'speaker_id','duration'])
+    dfd = df.drop(columns=['audio', 'speaker_id','duration'])
     return (df, dfd[:15], "all") #(df, df[:50])
 
 
@@ -62,14 +61,13 @@ def apply_filters(df,langname,gender):#,ageIS,ageFO):
         # filter by dialect
             
     if gender != "all":
-        df = df.filter(lambda x: x["gender"].startswith(gender))
+        df = df[df.gender.str.startswith(gender)]
 
     #if age != "all":
     #    df = df.filter(lambda x: x["age"] == age)
 
 
-    dfd = df.data.to_pandas()
-    dfd = dfd.drop(columns=['audio', 'speaker_id','duration'])
+    dfd = df.drop(columns=['audio', 'speaker_id','duration'])
     return (df,dfd[:15])
 
 
@@ -79,15 +77,14 @@ def f1(langname,ds):
     elif langname =="Faroese":
         lang_aligner = datas.a_f
 
-    maxdat=len(ds)
-
-    ds = ds.select([random.randint(maxdat-1)])
-    sound_path = ds['audio'][0]['path'] # audio 0 array is the audio data itself
-    transcript = ds['normalized_text'][0]
+    
+    ex = ds.sample()
+    sound_path = ex['audio'].iloc[0]['path']
+    transcript = ex['normalized_text'].iloc[0]
 
-    rec_info = f"{ds['audio_id'][0]}, {ds['gender'][0]}, {ds['age'][0]}"
+    rec_info = f"{ex['audio_id'].iloc[0]}, {ex['gender'].iloc[0]}, {ex['age'].iloc[0]}"
     if langname =="Faroese":
-        rec_info += f", {ds['dialect'][0]}"
+        rec_info += f", {ex['dialect'].iloc[0]}"
     return (graph.align_and_graph(sound_path,transcript,lang_aligner),sound_path,rec_info)